It is unclear in the U.S., and estimates vary widely. The number of undetected cases is estimated to range from 3 to 10 times the number officially recorded cases. I can't tell at all which is right. However, in China, Japan and Korea they have now done extensive random sample antibody tests, so they know approximately how many cases they missed. I think it is only around 2 times the cases they caught. The death rate for cases they know about in Japan is 5%. That is, 977 deaths divided by 18,934 cases. There have been few cases in recent weeks and most patients are recovered, so those are the final numbers. (For the first wave, anyway; there may be a second wave.) Assuming they only caught half the cases, the fatality rate was around ~2%. That number keeps coming up, so I think that's it. That is for Asian societies and Italy with many elderly people. It may apply to the U.S. because we have many obese people, who are as much at risk as normal-weight people in their 60s.
The Koreans tested extensively at the height of their epidemic. They later found few hidden cases with antibody tests. Their final numbers are 282 deaths divided by 12,904 cases = 2% case mortality. Claims that it is only 1% in Korea are wrong, I think. It might be ~1% now because there are so few cases they can effectively protect elderly people, so only feckless young people are getting sick. 30 to 50 per day. There are a few deaths per week.
In Japan, I think most patients are feckless middle aged nitwits who go to late night bars. The numbers are way up, to near 200. They say they will test all bars and karaoke joints. They better do it quickly!